Assignment: Study the Support Vector Machine Code - Implement IRIS Dataset (From SKLearn) on this given code i guess

# load libaries import numpy as np import math class SVM: def fit(self, data): # Our goal is to find out w ( Direction ) and b ( Interception ) for optimial hyperplane wx+b # for that, we'll try various w and b values b_step_size = 2 b_multiple = 5 # finding possiblity of hyperplane in left top, right top, left bottom and right bottom axes. axes = [[1, 1], [-1, 1], [-1, -1], [1, -1]] # let's find out maximum and minimum feature value features = np.concatenate((data[-1], data[1])) max_feature_value = np.max(features) min_feature_value = np.min(features) learning_rate = max_feature_value * .001 # Possible b values b_range = np.arange(-1 * (max_feature_value * b_step_size), max_feature_value * b_step_size, learning_rate * b_multiple) # this is just assumation for starting point. w_optimum = max_feature_value / 2 w = np.array([w_optimum, w_optimum]) optimized = False hyperplane_collection = {} while not optimized: for b in b_range: for section in axes: may_be_w = w * section suitable_hyperplane = True for one in data: for feature_data in data[one]: if one * (np.dot(may_be_w, feature_data) + b) < 1: # if any feature data is within the hyperplane margin, that's not the suitable hyperplane. suitable_hyperplane = False if suitable_hyperplane: # Calculate length of hyperplane direction hyperplane_norm = np.linalg.norm(may_be_w) # || w || # we need to keep track of these possible hyperplane. later, we'll choose which has minimum hyperplace norm as optimizal hyperplane. hyperplane_collection[hyperplane_norm] = [may_be_w, b] # Save w and b if w[0] < 0: optimized = True else: w = w - learning_rate # now sort by norm length norms = sorted([n for n in hyperplane_collection]) minimum_norm = norms[0] # This one is optimizal hyperplane optimal_hyperplane = hyperplane_collection[minimum_norm] self.w = optimal_hyperplane[0] self.b = optimal_hyperplane[1] # fit(data) def calculate_y(self, x, w, b, v): # w.x+b = v # w0x0 + w1x1 + b = v # x1 = (v - w0x0 - b)/w1 y = (v - w[0] * x - b) / w[1] return y def show_hyperplane(self, data): import matplotlib.pyplot as plt w = self.w b = self.b features = np.concatenate((data[-1], data[1])) x1 = np.max(features) y1 = self.calculate_y(x1, w, b, 0) x2 = np.min(features) y2 = self.calculate_y(x2, w, b, 0) figure = plt.figure() ax = figure.add_subplot(1, 1, 1) plt.scatter(data[-1][:, 0], data[-1][:, 1], marker="o", c="r") plt.scatter(data[1][:, 0], data[1][:, 1], marker="+", c="g") # This line should clearly separate male and female dataset ax.plot([x1, x2], [y1, y2], 'k') plt.show() def predict(self, features): w = self.w b = self.b classification = np.sign(np.dot(np.array(features), w) + b) return classification # Sample Data - (weight (kg), Height (cm), gender) sample_data = np.array([ [70, 175, 'male'], [60, 140, 'female'], [80, 185, 'male'], [75, 180, 'male'], [65, 150, 'female'], [70, 155, 'female'], [75, 160, 'female'], [85, 195, 'male'], [55, 170, 'female'], [65, 175, 'female'], ]) features = sample_data[:, :2].astype('float') target = sample_data[:, 2] # Assign -1 to female and +1 to male plusone = [] minusone = [] for i, gender in enumerate(target): if gender == 'male': plusone.append(features[i]) else: minusone.append(features[i]) data = {-1: np.array(minusone), 1: np.array(plusone)} classify = {-1: 'Female', 1: 'Male'} # Target - What is gender if weight is 50kg and height is 172cm ? guess = [[65, 180]] # Setup SVM Model model = SVM() # Train the Model model.fit(data) # Output prediction = model.predict(guess).astype('int') print("Gender is %s" % (classify[prediction[0]])) # Show optimial hyperplane. model.show_hyperplane(data)